Skip to main content
GET
/
api
/
core
/
models
Get available chat models
curl --request GET \
  --url https://api.example.com/api/core/models \
  --header 'Authorization: Bearer <token>'
[
  {
    "chat_model_name": "<string>",
    "label": "<string>",
    "description": {},
    "provider": "<string>",
    "icon_type": "icon",
    "icon": "<string>",
    "input_cost_per_token": 123,
    "category": "most_capable",
    "stability": "stable",
    "max_input_tokens": 128000,
    "region": "<string>",
    "default": false
  }
]

Authorizations

​
Authorization
string
header
required

Bearer token authentication. Format: 'Bearer '

Response

200 - application/json

Successful Response

​
chat_model_name
string
required
​
label
string
required
​
description
Description · object
required
​
provider
string
required
​
icon_type
enum<string>
required

Types of icons for the supported AI chat models. ICON: The provided icon will be a string that represents a font-awesome icon. NUMBER: The provided icon will be a string that will be displayed in the frontend, preferably a number (for example a '4o' for GPT-4o).

Available options:
icon,
number
​
icon
string
required
​
input_cost_per_token
number
required
​
category
enum<string>
required

The category of the chat model. MOST_CAPABLE: The model is one of the most capable in terms of performance for its provider. MOST_CAPABLE_NON_REASONING: The model is one of the most capable in terms of performance for its provider, but without reasoning capabilities. BALANCED: The model is balanced in terms of cost, speed and capability. FAST_AND_CHEAP: The model is cheaper and faster than the most capable models. LEGACY: The model is a legacy model that in most cases is slower and costs more than the most capable models. In some edge cases it might still perform better for specific tasks.

Available options:
most_capable,
most_capable_non_reasoning,
balanced,
fast_and_cheap,
legacy
​
stability
enum<string>
required
Available options:
stable,
experimental,
preview
​
max_input_tokens
integer
default:128000
​
region
string | null
​
default
boolean
default:false